POVERTY MEASUREMENT

FROM NOISE TO SIGNAL…

AND HOW THE MEDIA CAN HELP

João Pedro Azevedo

Global Poverty Practice
World Bank Group
WHY HAVE THIS TRAINING?
• TajStat has spent a long time without reporting poverty

• Poverty reduction is a very important policy objective for the Government of

Tajikistan

• Most people have a perception of what it means to be poor

• Most countries define monetary poverty in a very different way from the general
public perception

• The media is a critical stakeholder to both help improve the measurement of

poverty, and to promote poverty reduction as a priority

• The media is a critical actor to help move the debate from measurement to
action
WHAT IS POVERTY?
• Poverty is “pronounced deprivation in well-being”. The poor are those who do not
have enough income or consumption to put them above some adequate minimum
threshold.

• Poverty may also be tied to a specific type of consumption; for example, people could
be house-poor or food-poor or health-poor. These dimensions of poverty often can be
measured directly, for instance, by measuring malnutrition or literacy.

• The broadest approach to well-being (and poverty) focuses on the capability of the
individual to function in society. Poor people often lack key capabilities; they may
have inadequate income or education, or be in poor health, or feel powerless, or lack
political freedoms.
WHY IS IT IMPORTANT TO MEASURE POVERTY?
• To compare the poverty of different households and regions in the
country

• To compare poverty over time

• To compare poverty between countries

• To monitor poverty

• To define a poverty reduction strategy

• To evaluate the impact of a poverty reduction strategy

The answer to some of these questions will condition the properties

and choices of a poverty indicator.

4
DESIRABLE CHARACTERISTICS OF AN OFFICIAL
POVERTY MEASURE

• It must be understandable and easy to describe

• It must conform to a common sense notion of poverty

• It must fit the purpose for which it is being developed

• It must be technically solid

• It must be operationally viable

• It must be easily replicable

5
How is poverty measured?
 POVERTY MEASURES

Relative

Monetary

Absolute

Objective
Basic Needs
Approach
Poverty Non-monetary
Anthropometric
needs
Subjective Direct surveys

7
MEASURING MONETARY POVERTY
Three steps:

1. Estimate a poverty line: Absolute or relative

2. Chose a welfare aggregate: Income or Expenditure

3. Calculate aggregate the share of the population below this threshold (poverty line)

8
ABSOLUTE POVERTY
• Insufficient monetary resources to purchase the minimum consumption
basket socially accepted.

• The “poor” are those individuals whose expenditure (or income) are
below what is socially acceptable as a minimum living standard.

Non poor

Threshold

Poor

9
INTERNATIONAL BENCHMARK OF NATIONAL OFFICIAL
POVERTY LINES
NATIONAL OFFICIAL POVERTY LINES AND PRIVATE
CONSUMPTION PER CAPITA
• There is a clear positive relationship between
level of economic development and cost of
living, and the values of national official
poverty lines in PPP terms.

• A few countries are clearly outliers. Countries

such as Paraguay or Colombia before their
monetary poverty revision had a national
official poverty line much higher than we
could expect given is value of economic
development

• It is important to keep in mind that this

differences to not reflect only norms,
preferences and level of development.
Characteristics such as climate (heating and
clothing costs), geography (trade and
logistics), and demographics also play an
important role explaining some of this
differences.
Data from: Ravallion, Chen and Sangrula (2008)
Dollar a Day Revisited. Policy Research Working
Paper 4620.

1
 LIMITS OF MONETARY POVERTY MEASURES
• Implicit assumption: the distribution of resources within the household is
equitable

• Monetary poverty measures if the household has the possibility to purchase the
minimum consumption basket. It tells us nothing about what the household is actually
purchasing.

• Often non-poor household have caloric deficits

• The poverty line is socially determined, and it is country and time specific.

• Different societies can have different norms, and those norms can evolve over
time as income level, relative prices, technology, and preferences change.

• As consumption patterns change over time it is important to update the poverty line
(every 10 years).

1
NORMATIVE OR EMPIRICAL BASKETS?
• Nothing ensures that the items of a normative baskets will be available for
consumption

• There is no guarantee that these items have prices that can be estimated robustly

• Nothing ensures that the products from a normative food basket can be consumed by
a significant fraction of the population (products that are not part of the population
food habits).

• The normative baskets, use multiple criteria (caloric, protein intake, micro
nutrients, % protein of animal origin, etc.) end up being ad hoc baskets that
depend largely on the expert who built them. The criteria of transparency,
acceptance by consensus and reproducibility are hardly met in this case.

• An empirical basket is preferable

1
ESTIMATION OF THE FOOD POVERTY LINE
• Definition of an empirical basket and not a normative

• Select the list of most frequent products (criteria: frequency; shared of

expenditure; caloric contribution).

• Estimate the quantities purchased by the population of reference

• Convert such quantities in calories

• Sum calories and re-estimate the quantities

• Estimate unit values of the items using the survey (median values of the
reference population).

• Sum values (adjusted amounts * unit values) for food poverty line

1
THE NON-FOOD COMPONENT OF THE POVERTY LINE
Two methods:

• Estimated average Engel (or OLS) coefficient of the reference

population (most used method)

• The non-food component is that of households whose total costs is just

enough to buy the minimum food basket (Ravallion "austere" line)

1
FINAL CONSIDERATIONS
Clarity on the objective of the measure, transparency and simplicity are the
most critical aspects of this work
• Methodological guidelines
• Access to the survey data
• Clear calendar for the release of the numbers